Author: Vincent Diepeveen
Date: 06:00:00 11/18/99
Go up one level in this thread
On November 18, 1999 at 01:17:54, Michel Langeveld wrote: >On November 17, 1999 at 23:10:22, James Robertson wrote: > >>My current book format is just an array of hash codes that the program jumps >>through trying to find a match for the current position. There must be something >>faster, so I am wondering what other people do. If programmers could explain to >>me how their books are formatted, I would really appreciate it. >> >>Thanks in advance, >>James > >The format is OK in my opinion (Crafty uses the same format). In this way it's >possible to detect move-order-change, etc. I think if you sort the book you can >find fast position by searching binairy. In this way if you have 65536 position >you can find any book position by reading "only" 16 position from the book. This is true in my program Michel, but only if you store positions with moves, and not only positions. If you store positions you need to do it for every legal move. Therefore i store positions with moves in DIEP. Inside in diep i also have a redblack-tree that has the tournament book, which is probed before the big stupid book. Note that in ICCA journal it said my book was 200k, which is dead wrong. My book in Paderborn was 200 moves (tournament book). The big stupid book was about half a million positions with 1 or more moves each >Michel Langeveld
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.